Nonparametric Tests Of Association Research Articles

Validity, sensitivity and specificity of the mentation, behavior and mood subscale of the UPDRS

Objective: The unified Parkinson's disease rating scale (UPDRS) is the most widely used tool to rate the severity and the stage of Parkinson's disease (PD). However, the mentation, behavior and mood (MBM) subscale of the UPDRS has received little investigation regarding its validity and sensitivity. Three items of this subscale were compared to criterion tests to examine validity, sensitivity and specificity.Methods: Ninety-seven patients with idiopathic PD were assessed on the UPDRS. Scores on three items of the MBM subscale, intellectual impairment, thought disorder and depression, were compared to criterion tests, the telephone interview for cognition status (TICS), psychiatric assessment for psychosis and the geriatric depression scale (GDS). Non-parametric tests of association were performed to examine concurrent validity of the MBM items. The sensitivities, specificities and optimal cutoff scores for each MBM item were estimated by receiver operating characteristic (ROC) curve analysis.Results: The MBM items demonstrated low to moderate correlation with the criterion tests, and the sensitivity and specificity were not strong. Even using a score of 1.0 on the items of the MBM demonstrated a sensitivity/specificity of only 0.19/0.48 for intellectual impairment, 0.60/0.72 for thought disorder and 0.61/0.87 for depression. Using a more appropriate cutoff of 2.0 revealed sensitivities of 0.01, 0.38 and 0.13 respectively.Discussion: The MBM subscale items of intellectual impairment, thought disorder and depression are not appropriate for screening or diagnostic purposes. Tools such as the TICS and the GDS should be considered instead.

Telephone survey of private patients' views on continuity of care and registration with general practice in Ireland

BackgroundThe desire of patients for personal continuity of care with a General Practitioner (GP) has been well documented, but not within non-registered private patients in Ireland. This study set out to examine the attitudes and reported behaviours of private fee-paying patients towards continuity of GP care and universal registration for patients.MethodsCross-sectional telephone survey of 400 randomly chosen fee-paying patients living within County Dublin. There is no formal system of registration with a GP for these patients. Main outcomes were attendance of respondents at primary health care facilities and their attitudes towards continuity of care and registration with a GP. Data was analysed using descriptive statistics and using parametric and non-parametric tests of association. Pearson correlation was used to quantify the association between the described variables and attitudes towards continuity and registration with a GP. Variables showing significance at the 5% level were entered into multiple linear regression models.Results97% of respondents had seen a GP in the previous 5 years. The mean number of visits to the GP for respondents was 2.3 per annum. 89% of respondents had a regular GP and the mean length of time with their GP was 15.6 years. 96% preferred their personal medical care to be provided within one general practice. 16% of respondents had consulted a GP outside of their own practice in the previous year. They were more likely to be female, commute a longer distance to work or have poorer health status. 81% considered it important to be officially registered with a GP practice of their choice.ConclusionBoth personal and longitudinal continuity of care with a GP are important to private patients. Respondents who chose to visit GPs other than their regular GP were not easily characterised in this study and individual circumstances may lead to this behaviour. There is strong support for a system of universal patient registration within general practice.

Paraplegia: a postal survey of morbidity trends in India

Postal survey from August 2004 to May 2006. To ascertain the morbidity trends in individuals with paraplegia in India and to find its association with demographic characteristics. India. The questionnaire was mailed to the identified individuals (n=600) whose addresses were obtained from the medical records section of our hospital and by contacting non-government organizations (NGOs), working for individuals with paraplegia in various cities. The causes of morbidities surveyed were respiratory complications, use of catheter, pressure sores, spasticity, postural hypotension, pain and fractures. Data were analysed using nonparametric test of association (Goodman Kruskal Tau). A total of 276 (46%) individuals responded. Of all the morbidities studied, pain was the leading cause (57.2%) followed by spasticity (39.1%), pressure sore (28.3%), postural hypotension (10.1%), respiratory complications, and fractures (5.8%). We found significant associations between various morbidities and demographics and between morbidities themselves. The most common cause for morbidity was pain. Ambulation reduced the incidence of secondary complications.

Mental health in general practice: assessment of current state and future needs.

To determine the extent of mental health services provided in the community in one Irish health board area. To examine the influence of postgraduate mental health training of GPs on provision of mental health services. Questionnaire and focus group methods were employed to determine views on mental health service provision. Data analysis was with parametric and non-parametric tests of association including student's t and chi-squared tests. Thematic analysis of the focus groups was carried out. Twenty-five per cent of patients attending general practice have mental health problems and over 95% of these problems are dealt with in primary care. Only 32% of GPs had received postgraduate training in psychological therapies. GPs with postgraduate training in psychological therapies were more likely to estimate a higher proportion of their patient population with mental health problems and less likely to refer to psychiatric services. A need for support from other health care professionals in primary care was also identified. The majority of patients with mental health problems are treated in primary care. Further training of GPs and increased resources would improve mental health care in primary care and lead to fewer referrals to psychiatric services.

Rank-Sum Tests for Clustered Data

The Wilcoxon rank-sum test is widely used to test the equality of two populations, because it makes fewer distributional assumptions than parametric procedures such as the t-test. However, the Wilcoxon rank-sum test can be used only if data are independent. When data are clustered, tests based on generalized estimating equations (GEEs) that generalize the t-test have been proposed. Here we develop a rank-sum test that can be used when data are clustered. As an application, we use our rank-sum test to develop a nonparametric test of association between a genetic marker and a quantitative trait locus. We also give a rank-sum test for equivalence of three or more populations that generalizes the Kruskal–Wallis test to situations with clustered data. Unlike previous rank tests for clustered data, our proposal is valid when members of the same cluster belong to different groups, or when the correlation between cluster members differs across groups.

Nonparametric Tests of Association of Multiple Genes with Human Disease

The genetic basis of many common human diseases is expected to be highly heterogeneous, with multiple causative loci and multiple alleles at some of the causative loci. Analyzing the association of disease with one genetic marker at a time can have weak power, because of relatively small genetic effects and the need to correct for multiple testing. Testing the simultaneous effects of multiple markers by multivariate statistics might improve power, but they too will not be very powerful when there are many markers, because of the many degrees of freedom. To overcome some of the limitations of current statistical methods for case-control studies of candidate genes, we develop a new class of nonparametric statistics that can simultaneously test the association of multiple markers with disease, with only a single degree of freedom. Our approach, which is based on U-statistics, first measures a score over all markers for pairs of subjects and then compares the averages of these scores between cases and controls. Genetic scoring for a pair of subjects is measured by a "kernel" function, which we allow to be fairly general. However, we provide guidelines on how to choose a kernel for different types of genetic effects. Our global statistic has the advantage of having only one degree of freedom and achieves its greatest power advantage when the contrasts of average genotype scores between cases and controls are in the same direction across multiple markers. Simulations illustrate that our proposed methods have the anticipated type I-error rate and that they can be more powerful than standard methods. Application of our methods to a study of candidate genes for prostate cancer illustrates their potential merits, and offers guidelines for interpretation.

Developing R&D capacity in a primary care trust: use of the R&D culture index

There has traditionally been a low level of engagement of primary care practitioners with research. In the UK, primary care trusts (PCTs) now have some responsibility for the encouragement of research and development in primary care. The aim of this study was to assess the current level of research activity and capacity for research within a PCT. A questionnaire, incorporating a recently developed and validated research and development culture index, was sent to all 572 health care professionals and staff under the auspices of North Tyneside Research PCT. Data analysis used nonparametric tests of association including chi-squared, Mann–Whitney U and Spearman's rank order correlation. There was a 50.3 per cent response rate to the questionnaire. Groups more likely to show an increased capacity for research included those with postgraduate qualifications and those in post for the least time. General practitioners were less likely than other professional groupings to declare personal skills or aptitude for research. The two most important factors thought to contribute towards the development of a culture of R&D were having access to people to support development and changes in professional practice and having access to training and development opportunities. The use of the R&D culture index enabled groups to be identified that may be more research interested and can therefore be targeted to increase research capacity. The R&D culture index could be used by other PCTs wishing to define and develop research capacity in primary care.

Rank tests of association for exchangeable paired data

We describe two rank tests of association for paired exchangeable data motivated by the study of lifespans in twins. The pooled sample is ranked. The nonparametric test of association is based on R +, the sum of the smaller within-pair ranks. A second measure L + is the sum of within-pair rank products. Under the null hypothesis of within-pair independence, the two test statistics are approximately normally distributed. Expressions for the exact means and variances of R + and L + are given. We describe the power of these two statistics under a close alternative hypothesis to that of independence. Both the R + and L + tests indicate nonparametric statistical evidence of positive association of longevity in identical twins and a negligible relationship between the lifespans of fraternal twins listed in the Danish twin registry. The statistics are also applied to the analysis of a clinical trial studying the time to failure of ventilation tubes in children with bilateral otitis media.

Graphical Assessment of Dependence

Nonparametric tests of association have been around for a long time, and new ones continue to be proposed to cope with specific forms of association. However, graphs have the potential to assess a far richer class of bivariate dependence structures than any collection of tests. This article describes how chi-plots, used in conjunction with the usual scatterplot, provides a useful practical tool in this regard.

Graft survival in pediatric liver transplantation

Background/Purpose: Liver transplantation is standard therapy for children with a variety of liver diseases. The current shortage of organ donors has led to aggressive use of reduced or split grafts and living-related donors to provide timely liver transplants to these children. The purpose of this study is to examine the impact of these techniques on graft survival in children currently treated with liver transplantation. Methods: Data were obtained on all patients less than 21 years of age treated with isolated liver transplants performed after January 1, 1996 in an integrated statewide pediatric liver transplant program, which encompasses 2 high-volume centers. Nonparametric tests of association and life table analysis were used to analyze these data (SAS v 6.12). Results: One hundred twenty-three children received 147 grafts (62 at the University of Florida, 85 at the University of Miami). Fifty-two (36%) children were less than 1 year of age at time of transplant, and 80 (55%) were less than 2 years of age. Patient survival rate was identical in the 2 centers (1-year actuarial survival rate, 88.4% and 87.1%). Twenty-five (17%) grafts were reduced, 28 (19%) were split, 6 were from living donors (4%), and 88 (60%) were whole organs. One-year graft survival rate was 80% for whole grafts, 71.6% for reduced grafts, and 64.3% for split grafts (P =.06). Children who received whole organs (mean age, 6.1 years) were older than those who received segmental grafts (mean age, 2.5 years; P <.01). Multifactorial analysis suggested that patient age, gender, and use of the graft for retransplant did not influence graft survival, nor did the type of graft used influence patient survival. Conclusions: The survival rate of children after liver transplantation is excellent independent of graft type. Use of current techniques to split grafts between 2 recipients is associated with an increased graft loss and need for retransplantation. Improvement in graft survival of these organs could reduce the morbidity and cost of liver transplantation significantly in children. J Pediatr Surg 36:1205-1209. Copyright © 2001 by W.B. Saunders Company.

A developmental examination of children's understanding of task difficulty in the physical domain

The purpose of this study was to examine children's understanding of task difficulty in the physical domain using the developmental component of Nicholls’ theory of achievement motivation as a framework. Children (N = 144; 8 boys and 8 girls at each age from 5–13 years) enrolled in public schools in a Southern city were participants. They were shown one normative and two objective arrays with structured interview questions following each array. A Piagetian structural analysis of the children's responses revealed the same three levels of understanding of task difficulty that Nicholls’ previously identified: egocentric, objective, and normative. Further, a nonparametric test of association between ordinal variables showed a significant positive relationship between children's age and their levels of understanding of task difficulty. No significant gender differences emerged for children's understanding of task difficulty. Children's personal preferences for task difficulty levels (i.e., easy, moderate, difficult) were assessed for those who were identified as possessing a mature conception of task difficulty (i.e., rated at Level 3). Qualitative analysis revealed that approximately 30% of the children indicated that if given a choice they would prefer to play an easy game that did not provide an optimal challenge. These results provide partial evidence to support Nicholls’ (1989) claim that as children develop a mature understanding of ability some will be at risk of displaying maladaptive motivational patterns.

A Developmental Examination of Children's Understanding of Effort and Ability in the Physical and Academic Domains

The purpose of this study was to test one aspect of the developmental component of Nicholls' theory of achievement motivation. Specifically, we examined and contrasted children's levels of understanding of effort and ability in the physical and academic domains. Students (N = 144; 8 boys and 8 girls for each age group from the ages of 5–13 years) participated in two sessions; in each they viewed two films and responded to questions posed in a subsequent structured interview. The films showed two children applying unequal effort which resulted in two possible outcomes: (1) the children achieved the same score; or (2) the lazier child outscored the hard-working child. A developmental-structural analysis of the children's responses was conducted, and the same four levels of understanding of effort and ability identified by Nicholls (1978) emerged. In addition, a fifth level (i.e., Level 0) was included which represented children who were unable to identify the hard-working student in the films or whose responses did not address the effort-ability issue. A nonparametric test of association between ordinal variables revealed a positive and significant relationship between children's ages and their levels of understanding of ability in both contexts. Children's conceptions of ability were significantly correlated across domains. No gender differences were evident in children's understanding of effort and ability in either the physical or academic contexts.

Nonparametric Tests of Association between Survival Time and Continuously Measured Covariates: The Logit-Rank and Associated Procedures

O'Brien's logit-rank procedure (1978, Biometrics 34, 243-250) is shown to arise as a score test based on the partial likelihood for a proportional hazards model provided the covariate structure is suitably defined. Within this framework the asymptotic properties claimed by O'Brien can be readily deduced and can be seen to be valid under a more general model of censoring than that considered in his paper. More important, perhaps, it is now possible to make a more natural and interpretable generalization to the multiple regression problem than that suggested by O'Brien as a means of accounting for the effects of nuisance covariates. This can be achieved either by modelling or stratification. The proportional hazards framework is also helpful in that it enables us to recognize the logit-rank procedure as being one member of a class of contending procedures. One consequence of this is that the relative efficiencies of any two procedures can be readily evaluated using the results of Lagakos (1988, Biometrika 75, 156-160). Our own evaluations suggest that, for non-time-dependent covariates, a simplification of the logit-rank procedure, leading to considerable reduction in computational complexity, is to be preferred to the procedure originally outlined by O'Brien.

Cospeciation in Host-Parasite Assemblages: Comparative Analysis of Rates of Evolution and Timing of Cospeciation Events

-Documentation of widespread cospeciation in a host-parasite assemblage requires statistical evidence that the congruence observed between the host and parasite phylogenies exceeds that expected by chance. Although the validity of this test rests on the assumption of independence of the host and parasite phylogenies, this critical assumption may be violated in many tests of cospeciation. Herein, we emphasize the need for rigorous tests of cospeciation in host-parasite assemblages, and we show how estimates of genetic distance can be used to investigate relative rates of evolution and timing of phylogenesis in the hosts and parasites once widespread cospeciation has been documented for the assemblage. The method involves a nonparametric test of association between genetic-distance matrices for the hosts and their parasites. If the association is statistically significant, the relationship is examined in greater detail using bivariate analysis. We use an example from our studies of pocket gophers and chewing lice to illustrate how genetic distances can be used to explore relative rates of genetic change in the two groups and to investigate relative timing of cospeciation events in the assemblage. [Cospeciation; host-parasite coevolution; rates of evolution; genetic distances; pocket gophers; chewing lice.] Within the broad context of host-parasite coevolutionary theory, the inference of cospeciation is considered appropriate for a given host-parasite assemblage if the hosts and their parasites show identical patterns of phylogenetic differentiation. In practice, however, identical patterns of phylogenesis in hosts and their parasites are only rarely observed. Thus, in most studies of cospeciation, the investigator must accept a certain degree of discordance between host and parasite phylogenies and ask whether evidence of cospeciation is widespread within the assemblage. The search for widespread cospeciation usually involves a general assessment of congruence between host and parasite phylogenies or between host and parasite taxonomic boundaries or geographic distributions (e.g., Kethley and Johnston, 1975; Kim et al., 1975; Eveleigh and Amano, 1977; Brooks, 1979; Fain, 1979; Moss, 1979; 1 Present address: Department of Biological Sciences, Northern Illinois University, DeKalb, Illinois 60115. Radovsky, 1979; Hellenthal and Price, 1984). In comparisons of host and parasite genealogical trees, the inference of widespread cospeciation usually involves subjective appraisal of branching similarity, wherein a high level of congruence is assumed to reflect widespread cospeciation of hosts and their parasites, and a low level of congruence indicates that the majority of observed host-parasite associations cannot be attributed to parallel phylogenesis. Until recently (Humphries et al., 1986; Brooks, 1987; Simberloff, 1987; Page, 1989, 1990), few workers have quantified similarity between host and parasite phylogenies. Because the inference of widespread cospeciation is unwarranted if the observed similarity can be explained by chance, it is critical that investigations of cospeciation include assessments of the statistical significance of topological similarity of host and parasite phylogenetic trees. Even more fundamental to documentation of cospeciation is the logical and statistical requirement for independence of

Some Problems and Approaches in Avian Mate Choice

Some Problems and Approaches in Avian Mate Choice